Psi used in auto-logistic model

questions concerning analysis/theory using program PRESENCE

Psi used in auto-logistic model

Postby nathanvs » Tue Feb 14, 2017 5:05 pm

I have a pretty straightforward if technical question... What exactly does PRESENCE do to calculate the annual psi estimate used in the auto-logistic term? I've read both the PRESENCE help document, and the original Yackulic et al. 2012 & Eaton et al. 2014 papers & supplements that the model was created for AFAIK, and neither clearly states the formulae used.

The help file gives the example:

Code: Select all
Xsite4,t-1 = (Wsite3 * ψsite3,t-1 + Wsite5 * ψsite5,t-1) / (Wsite3 + Wsite5 )


Are the ψ in this case unconditional psi (modeled probability only) or conditional psi (taking into account the actual detection history at that site in that year)? The papers discuss that conditional can be used, but the above help file example reads like it's just the general probability of occupancy (i.e., the unconditional psi derived parameter). If this is conditional psi, how exactly is this calculated within PRESENCE?

Thank you for your help.
nathanvs
 
Posts: 6
Joined: Fri Nov 20, 2015 5:07 pm

Re: Psi used in auto-logistic model

Postby jhines » Tue Feb 14, 2017 8:31 pm

Sorry for the confusion. PRESENCE will use conditional occupancy if 'psi1' is in the design matrix for gam or eps. To use unconditional occupancy, enter 'upsi1' in the design matrix. I'll update the documentation.
jhines
 
Posts: 381
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: Psi used in auto-logistic model

Postby nathanvs » Wed Feb 15, 2017 10:52 pm

Thanks for the reply, that's quite helpful! I was also wondering, how is conditional occupancy calculated for psi1? Is this correct?

Detected: 1
Undetected: [derived annual psi] * (1-p) ^ [# surveys]

Thanks again,
Nathan
nathanvs
 
Posts: 6
Joined: Fri Nov 20, 2015 5:07 pm

Re: Psi used in auto-logistic model

Postby jhines » Thu Feb 16, 2017 2:03 pm

Conditional occupancy is computed as:

Code: Select all
               Psi * Prod(1-p(j))
Psi-c = ---------------------------------------
            1-Psi + Psi*Prod(1-p(j))


where Prod(1-p(j)) = (1-p(1))*(1-p(2))*(1-p(3))*...(1-p(K))
jhines
 
Posts: 381
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA


Return to analysis help

Who is online

Users browsing this forum: No registered users and 1 guest

cron